CENTER FOR DRUG EVALUATION AND
`RESEARCH
`
`
`
`
`
`APPLICATION NUMBER:
`202107Orig1s000
`
`SUMMARY REVIEW
`
`
`
`
`
`

`

`Division Director Review
`
`Summary Review for Regulatory Action
`
`17,2012
`F_b .
`M H. Parks, MD.
`
`Subject
`Division Director Summary Review
`NDA/BLA #
`202107
`
`(cross reference IND 76480
`Su n lement #
`
`Applicant Name
`Corcept Therapeutics, Inc.
`Date of Submission
`A .ril 18, 2011
`
`PDUFA Goal Date
`
`Feb
`
`17, 2012
`
`Proprietary Name /
`Established
`S ‘
`
`Name
`
`Korlym (mifepristone immediate-release tablet)
`
`Dosa_e Forms / Stren_ h
`
`300-m tablets
`
`
`
`Proposed Indication(s)
`
`To control hyperglycemia in adult patients with
`endogenous Cushing’s syndrome with T2DM or
`glucose intolerance who have failed surgery or are not
`
`Action/Recommended Action for Approval
`NME:
`
`1. Introduction
`
`Corcept Therapeutics has submitted this new drug application (NDA) under Section 505(b)(2)
`of the Federal Food, Drug and Cosmetic Act (FDCA) for the use of Korlym® (mifepristone) in
`the treatment of patients with endogenous Cushing’s syndrome who have failed surgery or are
`not candidates for surgery.
`
`Cushing’s syndrome is due to hypercortisolism and its clinical and metabolic consequences. It
`is broadly separated into exogenous and endogenous forms, the former due to exogenous
`glucocorticoid administration for varied medical conditions and the latter due to the body’s
`over production of cortisol. Endogenous Cushing’s syndrome is further divided into ACTH-
`dependent and ACTH-independent forms to distinguish between an extra-adrenal or intra—
`adrenal pathology. 1 As this application is only for the treatment of endogenous Cushing’s
`syndrome, the remainder of this memo will refer to Cushing’s syndrome with an
`understanding that it is specific to only the endogenous forms of this condition.
`
`Approximately 80—85% of Cushing’s syndrome are ACTH-dependent with 80% of these due
`to a pituitary tumor (Cushing’s disease) and 20% due to ectopic ACTH secretion from a non-
`pituitary tumor with the most prevalent ones being bronchial carcinoid and small cell lung
`
`1 Pivonello R et a1. Cushing’s Syndrome. Endocrinologv and Metabolism Clinics ofNorth America. 2008:
`37(1): 135-149.
`
`Page 1 of 23
`
`Reference ID: 3089695
`
`

`

`Division Director Review
`
`cancer although any tumor of neuroendocrine origin may produce ACTH.2 Of the
`approximate 15-20% of ACTH-independent Cushing’s syndrome, the majority are due to an
`adrenal tumor. Cushing’s syndrome is a rare disease with an incidence of 0.7 to 2.4 per
`million population per year. This application received orphan designation on July 5, 2007.
`
`The diagnosis of Cushing’s syndrome requires a multitude of laboratory and radiologic tests
`whose discussion extends beyond the scope of this memo. The objective of the laboratory
`tests is to demonstrate inappropriate and sustained hypercortisolism to distinguish these
`patients from conditions such as pseudo-Cushings, severe depression, or cyclical Cushing’s.
`Reliance on just clinical presentations is not possible or acceptable as patients’ presentations
`are highly variable and span a wide spectrum that includes textbook descriptions of buffalo
`hump, violaceous striae, hirsutism and facial plethora to more subtle signs of just diabetes and
`depression. The etiology of the syndrome may also influence the clinical presentation. For
`example, the age range of patients with ectopic ACTH syndromes is generally a decade older
`than those with Cushing’s disease with a lower female to male ratio. Patients with ectopic
`ACTH syndrome or adrenal cancers may also present with more severe signs and symptoms of
`hypercortisolism, and due to the underlying malignant nature of the tumor, these patients often
`have greater morbidity.
`
`Regardless of the etiology of Cushing’s syndrome, the treatment goal is the same and in all
`cases, if the underlying tumor can be located, surgical resection is the preferred initial therapy.
`Medical therapy may be initiated prior to surgery to control the hypermetabolic state and is
`often relied upon afterwards if surgery is unsuccessful or the tumor recurs. In some patients,
`radiation therapy and/or bilateral adrenalectomy are considered. The available medical
`therapies are limited and unapproved for Cushing’s syndrome.3 Their use has been based on
`the knowledge of their effects at inhibiting certain enzymes in the adrenal steroidogenesis
`pathway (e.g., ketoconazole or metyrapone) or limited inhibition of ACTH (e.g.,
`somatostatin). Mifepristone employs a different strategy in treating Cushing’s syndrome:
`blockade of the glucocorticoid II receptor (GR) to inhibit the actions downstream from this
`receptor. It also blocks the progesterone and androgen receptor, the former activity being the
`basis for its use in termination of pregnancy.
`
`
`2. Background
`
`
`There were two main challenges in the review of this application. The first was a scientific
`matter and the second was a regulatory/legal one.
`
`On the scientific note, the trial design to establish safety and effectiveness of Korlym for this
`indication was limited by 1) the underlying medical condition and 2) the pharmacologic action
`of the drug. Given the rarity and progressive nature of the condition with limited medical
`options, the type of trial design would have to be an uncontrolled and open-label design in a
`limited number of patients. Such a trial design in a small sample of patients complicates
`
`2 Alexandraki K and Grossman A. The ectopic ACTH syndrome. Rev Endocr Metab Disord. 2010; 11: 117-
`126.
`3 Mitotane is an exception but it has a limited indication in only patients with adrenal carcinomas
`
`Page 2 of 23
`
`Reference ID: 3089695
`
`

`

`Division Director Review
`
`attribution of effect and safety to drug. The mechanism of action of the drug presented another
`complexity as to the appropriate endpoint to evaluate effectiveness of Korlym. Just as the
`diagnosis of Cushing’s syndrome requires evidence of elevated cortisol levels, the treatment of
`these patients relies on a demonstration of reduced cortisol levels as a measure of response
`and/or success. Since the drug’s selective antagonism of the GR does not result in reduced
`cortisol levels, this biomarker was not of any utility for establishing efficacy and could not be
`employed as a measure for dose titration. Sections 6.0 and 7.0 of my memo delve further into
`the trial design and how the reviewers considered multiple lines of evidence to make a
`determination of safety and effectiveness.
`
`The regulatory and legal challenge of this application is because of the more controversial use
`of this active ingredient for medical termination of pregnancy in the approved formulation,
`Mifeprex®. Given as one-time lower doses than proposed in Cushing’s syndrome,
`mifepristone binds to the progesterone receptor (PR) to achieve pregnancy termination.
`Mifeprex, manufactured by Danco, was approved on September 28, 2000 under 21 CFR
`Subpart H and is available only through a restricted distribution program. With passage of the
`Food and Drug Administration Amendments Act (FDAAA) of 2007, a Risk Evaluation and
`Mitigation Strategy (REMS) with Elements to Assure Safe Use (ETASU) was applied to
`Mifeprex on June 8, 2011. Mifeprex is not distributed to or dispensed through retail
`pharmacies but is limited to specialty clinics and prescribed by physicians who have enrolled
`in a certification program. (Please see DRISK review for a full description of the Mifeprex
`REMS with ETASU).
`
`Prior to the submission of Korlym and throughout the NDA review, multiple internal meetings
`and discussions were held to determine if Korlym and its proposed indication met the
`regulatory requirements for a REMS with ETASU or if one would be necessary to maintain
`the integrity of Mifeprex’s REMS with ETASU.
`
`Dr. Dragos Roman in his cross-disclipine team leader (CDTL) memo has clearly outlined
`these discussions and the reader is also referred to memos written by DRISK reviewers, Drs.
`Robottom, LaCivita, and Karwoski, and meeting minutes prepared by Dr. Amy Egan for a
`meeting involving CDER Center Director and senior managers in OND, OSE, and ORP. On
`November 3, 2011, a CDER recommendation was made that given the rarity and seriousness
`of Cushing’s syndrome and the unique situation in which it would be used, a REMS with
`ETASU was not warranted. However, the applicant has agreed to establish a voluntary limited
`distribution system and a drug utilization study will be required postmarketing. Please see
`Section 13.0 for further discussions of the PMR for this application.
`
`
`3. CMC/Device
`
`
`CMC has recommended approval without any additional testing or studies required. Please
`see reviews of Drs. Ysern and Al-Hakim dated January 12, 2012.
`
`
`Page 3 of 23
`
`Reference ID: 3089695
`
`

`

`Division Director Review
`
`4. Nonclinical Pharmacology/Toxicology
`The applicant conducted several nonclinical studies to support the chronic use of mifepristone.
`These included safety pharmacology studies to evaluate potential of mifepristone to inhibit Ki
`channels, pharmacokinetic/ADME/and toxicokinetic studies, repeat-dose toxicity studies, in
`vitro genetic toxicology studies, and carcinogenicity studies. Published literature and studies
`conducted under approved NDA 20687 for use of mifepristone in pregnancy termination were
`also relied upon by the applicant as permitted under 505(b)(2). The three major metabolites of
`RU486 identified in humans were also present in mice, rats, dogs, and monkeys and were
`therefore adequately evaluated in the nonclinical program.
`
`Please see Dr. Patricia Brundage’s review dated January 19, 2012, for details of the nonclinical
`program supporting approval of this NDA. She and pharmacology/toxicology supervisor, Dr.
`Todd Bourcier, deem data acceptable in support of approval of mifepristone for Cushing’s
`syndrome provided labeling accurately reflects the nonclinical findings and their
`recommendations on use of the product. Dr. Bourcier’s memo dated February 7, 2012, also
`outlines the sufficient bridging data to Mifeprex® supporting reliance on FDA’s finding of
`safety and effectiveness for some aspects of this application. No postmarketing trials are being
`proposed by this discipline.
`
`Several of the safety findings identified reflect the pharmacology of mifepristone as an anti-
`glucocorticoid and anti-progesterogenic drug. The first of these effects is the basis for
`evaluating the use of mifepristone in the treatment of Cushing’s syndrome. Antagonism at the
`progesterone receptor is also included in the label and discussed in other sections of this memo
`with regard to the effect on fertility and pregnancy.
`
`Two hERG channel studies were performed of which one showed a concentration-related
`inhibition of potassium selective IKr current with mifepristone and its metabolites. A 12-
`month toxicity study in dogs also revealed a slight QTc prolongation in higher-dosed animals.
`These findings alongside the clinical tQT study support information on the potential QT
`prolongation effect of mifepristone in labeling with caution to be applied when used with
`drugs which might increase mifepristone drug exposures.
`
`
`5. Clinical Pharmacology/Biopharmaceutics
`Please see review of Drs. Jee Eun Lee and Jayabharathi Vaidyanathn dated January 13, 2012.
`Thirteen clinical pharmacology studies have been conducted by applicant and submitted to this
`NDA.
`
`Drug-drug interaction studies (DDI) were conducted with digoxin (P-gp substrate), alprazolam
`and simvastatin (CYP3A substrate), fluvastatin (insensitive CYP2C8/9 substrate), and
`cimetidine (mild CYP3A inhibitor). No DDI studies were conducted to address the effect of
`strong CYP3A4 inhibitors. The results from these studies are illustrated in the following
`figure:
`
`
`Page 4 of 23
`
`Reference ID: 3089695
`
`

`

`Effects of Mifepristone on Other Drugs
`
`
`
` Parameter Analyte Ratio [90% CI]
`
`Cmax
`AUC
`Cmax
`AUC
`Cmax
`
`AUC
`
`Cmax
`
`AUC
`
`Cmax
`
`AUC
`Cmax
`AUC
`
`Digox'n
`
`Alprazolam
`
`4.011 Alprazohm
`
`Simvastatin
`
`E:
`'I
`Ii
`E"
`If
`
`Ii
`
`H—1
`
`»—-—a
`
`1.64[ 1.50, 1.84]
`1.40[ 1.33, 1.47]
`0.81 [ 0.67, 0.99]
`1.81[ 1.52, 2.15]
`0.39[ 0.32, 0.47]
`
`0.76[ 0.64, 0.90]
`
`7.02 [ 5.27 , 9.34]
`
`10.40[ 8.11 , 1321 1
`
`Simvastatin acid
`
`»—-—>
`
`18.20 I 12.85 , 25.79]
`
`Fluvastatin
`
`E”
`
`m
`
`1—-—.
`
`15.70 | 1129 , 21.70]
`1.76[ 1.31 , 2.36]
`3.57[ 2.74, 4.65]
`
`r—I—I—I—I—I
`
`0
`
`5
`
`1o
`
`15
`
`20
`
`25
`
`Ratio
`
`*Simvastatin dose in multiple dosing regimen is 80 mg while 40 mg in single dosing regimen
`(Exposure was not normalized by dose)
`Figure 10. Forest plot for DDI with mtfeprlstone
`
`Given the significant increase in sensitive CYP3A substrate simvastatin, contraindications are
`proposed for simvastatin and lovastatin plus other CYP3A4 substrate drugs with narrow
`therapeutic indices.
`
`The applicant was asked to conduct a DDI with a potent CYP3A4 inhibitor. This was not
`done and the applicant instead provided 2 randomly-timed concentrations of mifepristone from
`one patient who was on concomitant ketoconazole th
`durin his artici ation in Stud
`400. This was not deemed acc
`
`table.
`
`Page 5 of23
`
`Reference ID: 3089695
`
`

`

`Division Director Review
`
`”(4)
`
`Afier further discussion it was decided that potential DDI with potent
`CYP3A4 inhibitors would be more appropriately discussed in Warnings and Precautions
`recommending against their use with mifepristone unless medically necessary. In addition the
`label will recommend limiting the dose of mifepristone to 300 mg daily if a strong CYP3A4
`inhibitor must be used concomitantly. A DDI study with ketoconazole will be required.
`Depending on the results of this study, updates to labeling may occur.
`
`Thorough QT Study
`A tQT study was conducted; please see the [RT review dated October 20, 2011. This study
`was a 14-day, multiple-dose, parallel treatment design enrolling 180 healthy male subjects but
`due to adverse events resulting in high discontinuation rates, data are limited out to Day 14.
`Mifepristone doses of 600 and 1800 mg were employed, the latter providing supratherapeutic
`exposures. A single oral dose of moxifloxacin 400 mg was used as a positive control.
`
`Overall, the study results were inconclusive because assay sensitivity was not established. The
`largest lower bound of the 2-sided 90% CI for the pbo—adjusted, baseline-corrected QTcI for
`moxifloxacin was < 5 ms; therefore, small changes in QTc interval cannot be excluded.
`Despite this shortcoming, the largest upper bounds of the 2-sided 90% CI for the mean
`difference between mifepristone 1800 mg and placebo did exceed 10 ms at several time points.
`No such finding was observed in the 600 mg dose group. This finding along with the one
`positive hERG channel study suggests a potential for QT prolongation associated with
`mifepristone use with increasing exposures.
`
`The proposed label will note that mifepristone and its metabolites block IKr, based on
`nonclinical data, and that Korlym prolongs QTc interval in a dose—related manner. More
`cautionary language is proposed with regard to using lowest effective dose. Since a DDI
`study has not been conducted with a potent CYP3A4 inhibitor, we will recommend limiting
`dose of Korlym to 300 mg if combined use with a potent inhibitor is necessary. The applicant
`will be required to conduct this DDI study with ketoconazole. Depending on the results,
`labeling may be revised accordingly.
`
`6. Clinical Microbiology
`
`Not applicable.
`
`7. ClinicalIStatistical-Efficacy
`
`Limitations of Clinical Development Program to be Considered
`One pivotal efficacy and safety trial was conducted by the applicant in support of this NDA.
`Published studies of mifepristone in Cushing’s syndrome were also submitted and summarized
`in Dr. Zemskova’s review but none of these studies was relied upon for the demonstration of
`efficacy of Korlym and are therefore not discussed in this memo.
`
`Page 6 of 23
`
`Reference ID: 3089695
`
`

`

`Division Director Review
`
`C1073-400 or Study 400 was a 24-week, open-label, uncontrolled trial that enrolled a total of
`50 patients with endogenous Cushing’s Syndrome. The stated objectives of the trial were to
`evaluate the safety and efficacy of mifepristone in the treatment of the signs and symptoms of
`endogenous Cushing’s syndrome. These are very broad objectives and in reality, only two
`endpoints were identified and patients were selected specifically to evaluate these endpoints:
`glycemic control and blood pressure.
`
`Before presenting the trial results, a discussion on several aspects of the trial design, endpoints,
`and patient population, and how they impact data interpretability will be necessary.
`
`Trial Design (Open-label and Uncontrolled)
`The open-label and uncontrolled nature of a trial can introduce confounders, biases, and
`limitations that are often mitigated through the conduct of a randomized, double-blind, and
`controlled trial. For this condition, a placebo control arm could not be employed because the
`progressive and serious nature of the condition would make it unethical to randomize any
`patient to placebo.
`
`An active-controlled trial to currently available therapies was not considered for several
`reasons. With exception for mitotane, which is approved for the treatment of inoperable
`adrenal cortical carcinoma of both functional and nonfunctional types, all medical therapies
`employed in practice for the treatment of Cushing’s syndrome are used off-label. The
`treatment regimens and efficacy of these other medical therapies have not been adequately
`assessed beyond case reports and anecdotal experience. In addition, these other drugs target a
`reduction in cortisol levels which cannot be achieved with Korlym by virtue of mifepristone’s
`mechanism of action. Selecting an appropriate and easily quantifiable endpoint that can
`compare the effects of the off-label therapies and Korlym could not be identified for a well-
`designed, active-controlled trial. Similarly, radiotherapy, which is also a treatment option in
`Cushing’s syndrome, would not be an appropriate active control given its variable success rate
`and time to demonstration of efficacy measured over the course of years.
`
`Untreated hypercortisolism in Cushing’s syndrome is progressive with little to no expectation
`of spontaneous improvement (e.g. the very rare instance of pituitary apoplexy in Cushing’s
`disease). For this reason, an uncontrolled trial of Korlym that could assess a clinically relevant
`efficacy endpoint might produce results which can be reasonably attributed to the drug.
`However, this limitation of the trial design must still be considered in the evaluation and
`conclusions made of the study results.
`
`Efficacy Endpoints
`As stated in the Introduction and Background sections of this memo, the mechanism of action
`of Korlym is antagonism of the GR preventing the downstream effects of cortisol on its
`receptor. Unlike other interventions targeting a reduction in cortisol levels, Korlym does not
`reduce serum cortisol and in some cases cortisol levels may increase. Furthermore, despite
`biochemical hypercortisolism, patients can become adrenally insufficient as a result of absent
`post-receptor activation.
`
`
`Page 7 of 23
`
`Reference ID: 3089695
`
`

`

`Division Director Review
`
`During the IND stage, FDA agreed with the applicant that demonstration of an effect on some
`other biochemical parameter will be accepted. The original protocol submission included very
`broad assessments of a composite clinical endpoint which was ultimately modified (with FDA
`input) to a demonstration on improvements in glycemic control and/or blood pressure defined
`as:
`
`
`1. The change in the area under the concentration-time curve for glucose (AUCglu) in the
`2-hr oral glucose tolerance test (oGTT) from baseline to Week 24
`2. A change from baseline to Week 24 in mean diastolic blood pressure (DBP)
`
`
`Secondary endpoints included a blinded assessment of selected signs and symptoms of
`Cushing’s syndrome as well as laboratory findings by a Data Review Board, body weight, use
`of concomitant medications for diabetes and hypertension, levels of HbA1c, systolic blood
`pressure, and photographs. There were exploratory efficacy variables assessed which will not
`be discussed in this memo. It should be noted that no hierarchical sequence for analysis was
`applied in the analysis of secondary endpoints. Although FDA did not object to this, FDA did
`note that control of Type 1 error for secondary endpoints would be important for consideration
`of labeling.
`
`The trial selected patients but did not randomize them into subgroups which would be
`evaluated specifically for one of the two primary endpoints. These subgroups are referred to
`as the Diabetes Mellitus (DM) and the Hypertension (HTN) cohorts.
`
`FDA has well-established efficacy criteria for therapies intended for the treatment of
`hyperglycemia in diabetes mellitus. The primary efficacy endpoint in both T1 and T2DM
`trials has been HbA1c which is a reliable measure of glycemic control and a surrogate for
`clinical benefits (e.g., microvascular complications). HbA1c was not selected as a primary
`efficacy endpoint in the Cushing Syndrome population because the clinical presentation of
`diabetes is variable in this condition and adjustments in anti-diabetic therapies are expected
`which would hinder the interpretation of results, especially in an uncontrolled study. As the
`DM cohort also included a few patients with glucose intolerance, a change in HbA1c from
`baseline might not be as reliable of a measure in these patients as they would have normal
`values at baseline. Reliance on a change in AUCglu during a 2-hr oGTT is a reasonable
`alternative assessment of glycemic response as it is under a controlled setting (unlike self-
`blood-glucose monitoring); is administered via a protocol; and is an objective laboratory
`measure. Nonetheless, results can be influenced by certain patient behaviors. Furthermore,
`unlike HbA1c, the clinical relevance of a reduction in AUCglu during an oGTT is unknown.
`Hence, the effect of Korlym on glycemic control will focus on both this primary efficacy
`measure supported by changes in HbA1c and anti-diabetic medications.
`
`Anti-hypertensive therapies have been approved based on mean changes in systolic and/or
`diastolic blood pressure. Hence, the endpoint of change in DBP is not unprecedented for drug
`approval. However, the effect of Korlym on blood pressure proved to be more difficult to
`demonstrate than anticipated and the results were obfuscated by the inclusion criteria, protocol
`violations, and use of certain anti-hypertensive medications. In retrospect, establishing
`efficacy of Korlym in Cushing’s syndrome based on blood pressure reduction should not have
`
`Page 8 of 23
`
`Reference ID: 3089695
`
`

`

`Division Director Review
`
`been considered a primary endpoint because the increased cortisol levels resulting from the
`drug’s mechanism of action may actually exacerbate hypertension secondary to
`mineralocorticoid receptor activation. Nonetheless, this memo will highlight these results
`from both an efficacy and safety perspective.
`
`Patient Population
`Given the rarity of this condition, the sample size in the pivotal trial was only 50 which is a
`limitation for evaluating efficacy and safety for chronic use but not unexpected for orphan
`indications. FDA has approved other therapies for rare disease with similar sample sizes
`(NDA for Increlex included 71 pediatric patients with severe Primary IGF-1 deficiency).
`
`It should be noted that despite the limited patient numbers in the pivotal trial, other clinical
`data of mifepristone in Cushing’s patients from published literature served as supportive
`evidence for efficacy and informed us in the design of the Phase 3 trial. None of these studies,
`which are summarized under Section 6.1.10, 7.7.2, and 9.1 of Dr. Zemskova’s review, will be
`included in labeling.
`
`Efficacy in Diabetes (DM) Cohort
`There were 29 out of 50 patients enrolled in Study 400 who were evaluated under the DM
`Cohort. Twenty-four (83%) had Cushing’s disease; 3 had ectopic ACTH and 2 had adrenal
`carcinoma. Twelve of the 24 patients with Cushing’s disease had prior radiation therapy
`whose data were reviewed separately by Dr. Zemskova to determine whether this previous
`treatment could account for any observed efficacy associated with Korlym. In her review, she
`pointed out that ACTH levels remained elevated in these patients despite radiation therapy.
`This would be evidence that radiation therapy was not successful and unlikely contributory to
`any efficacy observed in Study 400.
`
`Patients in this cohort underwent a 75-g oGTT at screening, on Day 1, Wks 6, 10, 16, and 24
`or on early termination visits. A patient was considered a responder if he/she had a 25% or
`more decrease in AUCglu at Wk 24 or early termination visit from baseline. A statistically
`significant reduction in AUCglu was defined by a responder analysis in which the lower bound
`of the 95% CI of this response rate had to exceed 20%. Approximately 60% of patients were
`responders and the lower bound of the 95% CI was 42%. From the cumulative distribution
`function curve provided by the applicant it is evident that the majority of patients had a
`reduction in AUCglu from baseline. The following table from Dr. Zemskova’s review
`summarizes the individual response for the 24 patients included in this analysis.
`
`
`Page 9 of 23
`
`Reference ID: 3089695
`
`

`

`Division Director Review
`
`
`
`
`In those patients who responded to treatment, a reduction in AUCglu was observed by Week 6
`in most patients and was sustained for the duration of treatment out to Week 24 (see Figure 3
`in Dr. Zemskova’s review).
`
`As stated above, AUCglu is not a standard efficacy endpoint for anti-diabetic medications and
`was accepted only for the unusual circumstances of evaluating glycemic control in Cushing’s
`patients treated with Korlym. However, the applicant also provided data on HbA1c reduction
`in 21 patients who had baseline and post-baseline values. The mean reduction from baseline
`in these patients was 1.14% (2-sided 95% CI: -1.56, -0.65; p=0.0001). This magnitude of
`reduction has been observed in currently approved anti-diabetic applications and considered to
`be clinically relevant. Dr. Zemskova further evaluated those patients with HbA1c levels above
`normal at baseline (6.5% - recall that trial enrolled patients with glucose intolerance or could
`have enrolled a diabetic patient with adequate control). In 14 patients with elevated HbA1c
`levels at baseline, all had a reduction from baseline including some with robust reductions of 2
`to 4% accompanied by reductions in anti-diabetic medications or doses.
`
`In conclusion, while this trial employed an untraditional measure of glycemic control and was
`uncontrolled, a correlation of AUCglu to clinically meaningful endpoints such as HbA1c
`reduction and dose reduction of antidiabetic medications could be established including
`several very robust effects (e.g., reduction from 10.4 to 6.0% in HbA1c level in one patient or
`
`Page 10 of 23
`
`Reference ID: 3089695
`
`

`

`Division Director Review
`
`~ 50% reduction in insulin requirements). These data constituted substantial evidence that
`Korlym would treat hyperglycemia associated with diabetes or glucose intolerance in
`Cushing’s syndrome. However, the observed effects should not be extrapolated beyond this
`patient population and it would be inappropriate to consider the use of Korlym in the
`management of diabetes unrelated to hypercortisolism due to Cushing’s syndrome. Labeling
`should indicate this as a Limitation of Use.
`
`Efficacy in Hypertension (HTN) Cohort
`Unlike the effect observed in the DM-cohort, the response rate in the HTN cohort was
`equivocal. Drs. Zemskova and Roman discuss some of the design flaws which might have
`contributed to the difficulty in establishing a robust effect and I will not reiterate them here. I
`believe that some aspect of the results reflect the pharmacologic effect of mifepristone.
`Hypertension in Cushing’s syndrome is partly due to high circulating levels of cortisol which
`can bind to the mineralocorticoid receptor. Acting like aldosterone, patients can present with
`hypokalemia and hypertension. Since mifepristone blocks the glucocorticoid receptor but does
`not cause a reduction in circulating cortisol levels, these patients are still prone to
`mineralocorticoid effects of hypercortisolism.
`
`Effects on Clinical Signs and Symptoms of Cushing’s Syndrome in Overall Cohort
`It should be noted that no plan was submitted to FDA for review by the Study Endpoints and
`Labeling of Drugs (SEALD) review team with respect to patient reported outcome measures.
`FDA reviewers have determined that while these endpoints should be evaluated in a clinical
`trial, the limitations of the assessments in an open-label, uncontrolled trial should preclude any
`quantitative statements in labeling.
`
` A
`
` Data Review Board (DRB) comprised of 3 experts on Cushing’s syndrome performed a
`review of 8 categories of clinical parameters to evaluate whether a patient’s signs and
`symptoms of Cushing’s had changed. These categories included:
`
`
`1. assessment of glucose homeostasis
`2. assessment of blood pressure
`3. assessment of lipids
`4. changes in weight and body composition
`5. clinical scoring and appearance (e.g., acne, hirsutism (in women only), Cushingoid
`appearance)
`6. strength assessments
`7. psychiatric and quality of life assessments
`8. metabolic bone assessments
`
`
`The DRB reviewed adverse events, concomitant medication data, and all efficacy assessments
`obtained at baseline, Weeks 6, 10, 16 and 24 or end of treatment, and at the follow-up visit.
`Baseline and follow-up evaluations were identified, but data from other visits were reviewed in
`a blinded fashion with respect to visit. The DRB did not know the dose of drug or the
`sequence in which the visit occurred. It should be noted that while the DRB reviewers are
`blinded to the sequence of visits, some of the assessments at each visit were evaluated by a
`
`Page 11 of 23
`
`Reference ID: 3089695
`
`

`

`Division Director Review
`
`clinical investigator who was NOT blinded. After reviewing the data, each DRB member
`assigned an overall score for each visit as follows:
`
`
` -1: worse than baseline
` 0: unchanged from baseline
`+1: clinically significant improvement
`
`
`The median of the 3 scores was calculated and a score of +1 was considered evidence of
`clinical improvement. A responder was defined as a subject whose median score was +1 at
`any visit after the baseline visit. Based on the applicant’s definition of responders, they report
`that 87% of patients (40/46) in the mITT population had a clinical improvement at any point in
`time and deemed the findings statistically significant based on a calculated 95% CI yielding a
`lower bound of > 30%, an arbitrary cut point considered by applicant as adequate to account
`for variability in response.
`
`Drs. Zemskova and Choudury appropriately point out the limitations of this endpoint
`assessment. The open-label nature of the trial is always problematic in evaluating subjective
`measures such as quality of life where patient reports may be perceptions based on
`expectations of clinical improvements or side effects based on knowledge that he/she is
`receiving an investigational agent. This is further compounded by the absence of a control
`group for comparison of response for the less subjective measures. In addition, declaring a
`patient as a responder at any visit also allows the applicant many opportunities for concluding
`success on this endpoint.
`
`Finally, it is not clear how the reviewers ranked the clinical relevance of the 8 clinical
`parameters in their scoring. The form for this scoring is provided below.
`
`
`Page 12 of 23
`
`Reference ID: 3089695
`
`

`

`Corcept Therapeutics
`
`Data Revnew Board Evaluation
`
`Study: C1073-400
`
`
`
`
`Scoring:
`
`-l = Worse than Baseline
`
`
`
`
`[:I NlA-No dam
`
`'1 N/A- No data
`
`
`
`
`
`+0 = Unchanged from Baseline
`=Clinicall
`si-
`lficant irn rovement from
`
`IRESULT.
`Imeclurkone Won '0»: box rm}:
`
`
`
`V'S'TA
`
`El
`
`
`
`
`a.
`
`
`VlSl'] D m +1
`Cl
`
`Reviewer Signature:
`
`
`
`
`month
`
`year
`
`There is no breakdown of how response in categories such as blood pressure or metabolic
`parameters might have conhibuted to the scores or if there was worsening in one and
`improvement in another, how these components wer